Hinglish captions — you speak Hindi, the captions come out in Roman

How to add Hinglish subtitles to a video

Four steps, start to finish, in a browser tab. Nothing to install, no account, and the video never leaves your device — so a forty-minute podcast starts captioning as fast as a thirty-second reel, with no upload to sit through first.

1. 1

   Open the editor and add your video. Go to aicaptioncut.com and drop your MP4, MOV or WebM onto the page, or press "Choose a video". The file is read straight off your disk — there is no upload, so there is no file-size limit and no waiting.

2. 2

   Set the language to "Hinglish (Indian)". This is the step every other tool is missing. In the language dropdown choose "Hinglish (Indian)", not "Hindi" — "Hindi" gives you Devanagari. Hinglish runs a Whisper model fine-tuned on Indian-accented Hindi that writes Roman letters directly, and anything that still comes back in Devanagari is romanised on your device before it reaches the caption list.

3. 3

   Press Generate captions. The model (about 80 MB) downloads once and is cached, then transcribes on your own CPU. "नमस्ते दोस्तों, आज एक नई बात" arrives as "namaste doston, aaj ek nayi baat", timed to the word — so karaoke highlighting and word-by-word animations land on the syllable being spoken.

4. 4

   Fix any spelling, style it, export. Tap any caption in the list to retype a word; the timing does not move. Then pick a caption style, adjust font, size, colour and position, and export an MP4 with the captions burned in — or save the subtitles as .srt or .vtt for another editor.

Captioning code-switched Hindi and English

Real Hinglish speech is not Hindi with a few English words dropped in — it switches mid-sentence, and a sentence like "matlab yaar the whole point hi yahi hai" is normal. That breaks caption tools in a specific way: a Hindi model transliterates the English words phonetically ("व्होल पॉइंट"), and an English model mishears the Hindi entirely.

Choosing "Hinglish (Indian)" is what avoids both. The checkpoint was trained on roughly 550 hours of exactly this speech — Indian-accented, code-switched, noisy — so English words stay spelled as English words and the Hindi around them comes out in Roman. You do not set a second language, and you do not mark where the switches happen.

One thing worth knowing before you start: there is no single correct spelling of Hinglish. People write "pyaar" and "pyar", "kyun" and "kyu". We pick one consistent spelling, and every word stays editable — most creators change two or three in a whole reel.

What you actually get

Pick "Hinglish (Indian)" as the spoken language and press generate. "नमस्ते दोस्तों, आज एक नई बात" comes back as "namaste doston, aaj ek nayi baat" — Roman letters, casual spelling, the way creators type. No copy-paste into a transliteration site, no retyping fifty caption lines by hand.

The timing survives the conversion. Each romanised word keeps the timestamp of the Hindi word it came from, so karaoke highlighting still lands on the syllable being spoken and word-by-word pop animations still hit on the beat.

Why other tools cannot do this

Speech-to-text models write Hindi in Devanagari, because that is the script the training data uses. Cloud caption tools hand you that output unchanged. The romanisation step is a separate piece of work, and it is one nobody bothered to build for Hindi creators.

Ours runs entirely in your browser — a phonetic transliterator, not an API call. That is also why it costs nothing and why your video is not uploaded to run it.

Honest about the spelling

There is no single correct way to spell Hinglish. People write "pyaar" and "pyar", "kyun" and "kyu", "mein" and "me". We pick one natural, consistent spelling — it will not always match your taste, and the occasional uncommon word comes out heavy ("ajanbee" rather than "ajnabi").

So every word stays editable. Tap a caption in the list and retype it; the timing does not move. Most creators fix two or three words in a whole reel and ship it.

Free, and no length cap

Caption a 30-second reel or a 40-minute podcast — there is no limit on length, file size or how many videos you do in a month. The reason is architectural, not generous. Cloud caption tools rent GPUs by the minute, so every minute of your video costs them money — which is why their plans meter you by video length, upload size or credits. AI Caption Cut renders on your own device, so a longer video costs us nothing. There is no meter to run.

AI Caption Cut is free to use, with unlimited exports at 480p carrying a small watermark. Removing the watermark is a one-off payment for that one video — 720p, or full 1080p at 60fps — or an Unlimited Pass that covers every export for thirty days. Subtitle files are free either way. It is charged in your own currency at checkout, and there is no subscription and no account.

Questions

Related free tools